THE EXPORT-IMPORT Bank is fighting for its life on Capitol Hill. To both advocates and opponents, one key bone of contention is jobs — specifically, how many does Ex-Im add, subtract or multiply? The pro-Ex-Im forces tout the figure of 205,000 jobs “supported” in fiscal 2013 by the agency’s loan guarantees and other financing for U.S. sales abroad.

Prior to a recent hearing before the House Financial Services Committee, a pro-bank group gave to each member of the panel a card on which was printed a summary of the jobs impact for his or her congressional district. To take one example, the card for Rep. John K. Delaney (D-Md.) claimed that, between fiscal 2007 and 2013, Ex-Im had “supported” $379.3 million worth of exports and 2,423 jobs at 15 firms in his district, which stretches from Gaithersburg to the Western Maryland panhandle.

It’s eminently plausible that Ex-Im helps maintain a level playing field for big U.S. firms in the heavily politicized and subsidized global market for jets, power plants and other capital goods. But when it comes to jobs, well, just how rigorous are the estimates, really? Congress ordered a study of that very question when it last reauthorized Ex-Im in 2012. In May 2013, the Government Accountability Office (GAO) produced its verdict: Meh.

GAO noted that Ex-Im must speak vaguely of “jobs supported,” rather than concretely of jobs created, since its methodology cannot really distinguish between new employment and retained employment. To get a number for “jobs supported,” which includes both a given firm and that firm’s suppliers, Ex-Im multiplies the dollar amount of exports it finances in each industry by a “jobs ratio” (calculated by the Bureau of Labor Statisics).

On average, Ex-Im calculates, a billion dollars of Ex-Im-backed exports supports 6,390 U.S. jobs; ergo, the $379.3 million exported by firms in Mr. Delaney’s district supported 2,423. GAO pointed out, additionally, that these figures do not differentiate between full-time and part-time work and, crucially, provide no information about what might have happened to employment at the firms in question, or others, if the resources marshalled by Ex-Im had flowed elsewhere in the economy.

GAO found nothing fraudulent about any of this, nor do we. The watchdog agency simply noted the rather crucial assumptions and limitations embedded in Ex-Im’s methodology and urged the bank to be more transparent about them — because “Congressional and public stakeholders may not fully understand what the jobs number that Ex-Im reports represents and the extent to which Ex-Im’s financing may have affected U.S. employment.” Ex-Im promised to lay bare its methodology on its Web site and annual report; the four paragraphs you’ll find there do indeed provide a bare-bones summary of the methodology, but not its limitations.

Perhaps because of space constraints, there wasn’t even that much explanation on the cards for Mr. Delaney and other members of Congress, though lobbyists did find room for the claim that Ex-Im helps “increase American jobs.” We hope our explanation helps put that assertion in context.